A link between renewable energy, globalisation and carbon emission? Evidence from a disaggregate analysis with policy insights

This study unveils the question of how renewable energy, non-renewable energy, globalisation, and total factor 6 productivity affect the carbon dioxide (CO 2 ) at the aggregate and disaggregate levels (CO 2 from oil, coal and gas) in 7 case of top ten carbon emitters developing economies over the period 1991-2016. To achieve the above objective, we 8 apply various panel unit, cointegration and causality tests. We also implement a Pooled Mean Group estimator 9 technique to find the long-term coefficients. Findings from panel cointegration tests show that there exists a significant 10 long-run relationship between renewable energy, non-renewable energy, globalisation, total factor productivity and 11 CO 2 . Moreover, findings derived from PMG infers that renewable energy consumption has a negative and significant 12 impact on CO 2 while non-renewable energy consumption significantly increases the CO 2 at aggregate and 13 disaggregate level. Further, our results show that total factor productivity increases the CO 2 emissions whereas 14 globalisation decreases it. From the policy point of view, our findings recommend that CO 2 in sample countries can 15 be reduced through promoting low carbon technology, and globalisation. Moreover, our findings propose to encourage 16 renewable energy installation and drafting comprehensive policies. statistically significant and positive. A 1% increase in coal consumption increases carbon emissions from coal by 0.90%.


39
The growth of the economy and the progress of industrialisation are resulting in massive amounts of fossil-fuel energy 40 usage. In recent years, globalised economies' economic and non-economic activities are based mainly on energy inputs 41 that simultaneously lead to energy security and sustainable development (IEA, 2017;BP Global, 2018). As a result, developing and developed countries it has lower energy intensity ratio. There is also need to identify different sources 73 of emission which vary across countries because the dynamic relationship with related impact factors also differs with 74 respect to sources of emission (Ahmad et al., 2016;Nain et al., 2017).   In Figure 2, we plot the share of different fossil-fuel like coal, natural gas and oil to total carbon emission of top ten 78 carbon emitters developing countries. From Figure 2, we visualize that there has been a substantial variation across 79 these countries in terms of energy and emission sources. For example, China, India, and South Africa heavily rely on 80 coal consumption, thereby having the largest share of CO2 emission.

81
Furthermore, renewable energy is a key component of handling the problem energy-security and reduction in 82 GHGs emissions. In addition, "it tells about non-exhaustive source of energy that should be increased for long-term 83 sustainability (Bhat, 2018)". According to existing studies, the government's initiative in recent years has resulted in a 84 significant decrease in the cost of renewable energy technology, which has evolved in tandem with the increase in energy

99
Despite the vital role of globalisation, TFP, renewable and non-renewable energy (NREC) demand and GHGs 100 emission, studies on the link between environment and its influencing macroeconomic factors are scanty. Hence, "there 101 is a need for close investigation of the relationship between environment and its influencing macroeconomic factors to 102 design a nuanced energy and environmental policy". Further, given the position of globalisation and technological 103 progress in the existing literature, the current study bridges this research gap by investigating the impact of globalisation, 104 TFP, renewable and non-renewable on the different carbon emission sources (or disaggregate level). In the global level,

105
we consider the sample of top ten emitting nations which is prime importance in international negotiation on climate 106 change. To the best of authors' knowledge, none of the previous studies examined the impact of globalisation, TFP, 107 renewable and non-renewable on carbon emission at the disaggregate level (emission from coal, gas and oil) in a panel 108 data framework in the top ten carbon-emitting countries among developing nations. In one of Ertugrul et al., (2016) 109 studies, he showed the impact of energy consumption on carbon emissions in top carbon emitters by taking aggregate 110 level into account. As a result, this study adds to the research on the carbon-influencing macroeconomic factors nexus 111 in the following ways. To begin, this work differs from previous works in that it uses TFP as a proxy for economic 112 growth to evaluate the role of total productivity in carbon emissions. Second, we explore long-run relationships and 113 elasticities using the advanced panel data model, i.e., pooled mean group (PMG). Because most cross-country studies 114 neglect the issue of cross-sectional dependency in the error term which lead to biased results. This problem is critical 115 from the perspective of global economic coordination on "climate change and voluntary carbon emission reduction".

116
Third, we have used a unique dataset of emission from coal, gas and oil-related to the top ten developing countries at 117 disaggregated levels which have the largest potential for reduction in emissions. Finally, we developed a robust 118 technique of long-term impact that incorporates both the cross and time dimensions of the data point, resulting in a 119 considerable improvement in estimation over studies that exclusively use the time series method.

120
The remaining part is assembled as follows: literature review section supply assessment of relevant studies.

121
The data and methodology part delineate the empirical modeling, data collection and methods of estimation. The    (2017). Given the role of renewable energy 133 consumption in recent years of government mission to achieve the full potential production of renewable energy, the 134 recent studies have distinctly studied the effect of renewable energy consumption along with non-renewable energy 135 consumption on economic growth and CO2 emission.

141
Furthermore, only a few researchers have looked into the impact of globalisation on CO2 emissions and energy 142 consumption, and various proxies of globalisation have been used as indicators of globalization, i.e., trade openness.

143
There are no clear-cut conclusions (or mixed ones) in terms of the dominance of size or the composition influence of 144 trade, there are no clear-cut conclusions (or mixed ones) (Cole, 2006;Copeland, and Taylor, 2004;Antweiler et al., and trade, however the evidence was inconclusive (Hossain, 2012;Shahbaz et al., 2013aShahbaz et al., , 2013bShahbaz et al., 2014a; 147 Nasreen and Anwer, 2014).

148
[Insert Table 1] 149 [Insert Table 2] 150 151 Existing studies have been divided into two portions to maintain the relevancy of the empirical investigations., (i) 152 studies based on a link between CO2 emission and renewable energy consumption is are given in Table 1; (ii) literature 153 on the relationship between globalisation, energy consumption and carbon emission (an indicator of environmental 154 quality) are reported in Table 2. Table 1 shows that no single study has come to the conclusion that increasing renewable       or not. As mentioned previously, these unit root tests produce an accurate outcome by using dynamic autoregressive 207 coefficient, which allows for heterogeneity across the sample countries and identify the order of integration of variables 208 very suitably. The order of integration either I(0) or I(1), is found through testing variables at levels and if it is not 209 stationary then we will proceed to apply unit root test at their first differences, this indicates that all the variables are 210 non-stationary at the level and stationary at first difference. This allows us to proceed to test for the existence of a long-211 run relationship (cointegration) among variables for the model (1-4).

219
rho-statistic and ADF-statistic. The second set comprises three group statistics known as between-dimension 6 includes 220 rho-statistic, ADF-statistic and PP-statistic. Null of no cointegration is tested against the alternate hypothesis that there 221 is cointegration among the variables. We have also used Kao panel cointegration test developed by (Kao C, 1999;1990).

222
On the first stage of regressors, these Kao panel cointegration techniques include homogeneous coefficients across all 223 units which follow a similar procedure as Pedroni cointegration techniques.

238
( ) Where z denotes the dependent variables (emissions; total, from coal, oil and gas), represents the error correction

278
( 1999) and Pesaran and Smith (1995) to explore the long term impact and short-term dynamics. The results of these 279 models are showed in Table 8. The coefficient of model 1 shows that increase in the NREC stimulates the level of carbon

291
The empirical results in Table 7 also imply that increase in the consumption of renewable source of energy

319
The relationship between non-renewable energy (coal consumption) and CO2 emissions from coal is found to be 320 statistically significant and positive. A 1% increase in coal consumption increases carbon emissions from coal by 0.90%.

361
The short-run results are shown in the lower part of Table 7. As can be seen, the effect of non-renewable, total factor 362 productivity, and globalisation are statistically insignificant at the aggregated and disaggregated levels. However, the 363 effect of renewable energy source is found to be significant but positive on carbon emissions in the (Model 2) of short-364 run analysis 7 . The error correction explains the adjustment speed towards a long-run path from short-run disequilibrium 365 at the aggregated and disaggregated level. It can be observed that ECTt-1 is statistically significant and negative at 5%

381
In order to perform the panel causality among CO2 emissions, renewable energy, non-renewable energy, total factor 382 productivity and globalisation in the top ten carbon emitters among the developing countries at the aggregate and 383 disaggregate levels, the Dumitrescu and Hurlin (2012) test is applied. This technique is considered to allow coefficient 384 to vary across cross sections and consider heterogeneity and cross-sectional dependence. We have used the first 385 difference series because the examined variables in the model should be stationary to run this test. LnREC to LnCO2coal. However, the causality between LnTFP, and LnCO2 and LnCO2coal is found to be bidirectional,

389
this indicate that total factor productivity causes total carbon emissions as well as carbon emissions generated from coal in the top carbon emitter's countries (model 2). Moreover, the consumption of oil also causes LnCO2 emissions from oil 391 consumption. Hence, we can conclude that energy consumption at aggregated and disaggregated levels causes 392 environmental pollution, whereas Globalisation mitigates while total factor productivity stimulates carbon emissions in 393 developing countries.